Sequence analysis
The sequence analysis was done using the OBITools package (Boyer et al., 2016). Briefly, forward and reverse reads were assembled usingilluminapairedend program before assigning the assembled sequences to the different samples using the ngsfilter . The original dataset was split into several files using the obisplitprogram, one file per sample and primer pair. Sequences corresponding most likely to PCR or sequencing errors were discarded. The discarded sequences were shorter than 20 bp, or occurring less than 10 times per sample, or labeled as ”internal” by the obiclean program. The taxonomic assignment was carried out with the ecotag program, using a curated reference database (Prié et al. 2020) or a database extracted from the EMBL release 138. Only sequences with a similarity higher than 98% with one of the databases were kept. All sequences with a frequency of occurrence below 0.001 were considered as potential tag-jumps (Schnell, Bohmann, & Gilbert, 2015) and discarded.